LPAS speech coder using vector quantized, multi-codebook, multi-tap pitch predictor and optimized ternary source excitation codebook derivation

• US 6,014,618 A
• Filed: 08/06/1998
• Issued: 01/11/2000
• Est. Priority Date: 08/06/1998
• Status: Expired due to Term

• Current Assignee: Tellabs Operations Incorporated (Infinera Corp.)
• Original Assignee: DSP Software Engineering, Inc. (Infinera Corp.)
• Inventors: Patel, Jayesh S., Kolb, Douglas E.
• Primary Examiner(s): Hudspeth, David R.
• Assistant Examiner(s): MCFADDEN, SUSAN IRIS
• Application Number:
  US09/130,688
• Time in Patent Office: 523 Days
• Field of Search: 704/219, 704/226, 704/222, 704/223, 704/207
• US Class Current: 704/207
• CPC Class Codes: G10L 19/08 Determination or coding of the excitation function; Determination or coding of the long-term prediction parameters
• CPC Class Codes: G10L 19/09 Long term prediction, i.e. removing periodical redundancies, e.g. by using adaptive codebook or pitch predictor
• CPC Class Codes: G10L 2019/0005 Multi-stage vector quantisation
• CPC Class Codes: G10L 2019/0011 Long term prediction filters, i.e. pitch estimation

Abstract

A method and apparatus for reducing the complexity of linear prediction analysis-by-synthesis (LPAS) speech coders. The method and apparatus include product code vector quantization (PCVQ) of multi-tap pitch predictor coefficients, which reduces the search and quantization complexity of an adaptive codebook. Further included is a procedure for generating and selecting code vectors consisting of ternary (1,0,-1) values, for optimizing a fixed codebook. Serial optimization of the adaptive codebook first and then the fixed codebook, produces a low complexity LPAS speech coder of the present invention.

Claims

1. In a system having a working memory and a digital processor, a method for encoding speech signals comprising the steps of:

providing an encoder executable in working memory by the digital process, the encoder including (a) a pitch predictor and (b) a source excitation codebook, the pitch predictor for removing certain redundancies in a subject speech signal, the pitch predictor having various parameters, and being a multi-tap pitch predictor utilizing a codebook subdivided into at least a first vector codebook and a second vector codebook, the source excitation codebook for indicating pulses in the subject speech signal;

vector quantizing the pitch predictor parameters such that computational complexity and memory requirements of the encoder are reduced, said vector quantizing employing product code vector quantization; and

in the source excitation codebook, deriving ternary values (1,-1,0) to indicate pulses of the subject speech signal, such that computational complexity of the encoder is further reduced.

2. A method as claimed in claim 1 wherein the step of providing an encoder includes providing a linear-predictive analysis-by-synthesis speech coder.

3. A method as claimed in claim 1 further comprising the step of sequentially searching the first and second vector codebooks.

4. A method as claimed in claim 1, wherein the step of providing an encoder including the source excitation codebook includes considering non-contiguous positions for each pulse, such that computational complexity is reduced.

5. A method as claimed in claim 1 further comprising the step of sequentially optimizing the pitch predictor and the source excitation codebook.

6. In a system having a working memory and a digital processor, apparatus for encoding speech signals comprising:

(a) a multi-tap pitch predictor for removing certain redundancies in a subject speech signal, the multi-tap pitch predictor having vector quantized parameters such that computational complexity and memory requirements of the apparatus are reduced, the multi-tap pitch predictor having a codebook subdivided into at least a first and a second vector codebook;

(b) a source excitation codebook coupled to receive speech signals from the pitch predictor, the source excitation codebook for indicating pulses in the subject speech signal, the codebook employing ternary values (1,0,-1) which are derived to indicate the pulses, such that computational complexity is further reduced.

7. Apparatus as claimed in claim 6 wherein the pitch predictor parameters are product code vector quantized.

8. Apparatus as claimed in claim 6 wherein the apparatus is a linear-predictive analysis-by-synthesis speech coder.

9. Apparatus as claimed in claim 6, wherein the first and second vector codebooks of the pitch predictor are sequentially searched.

10. Apparatus as claimed in claim 6 wherein the source excitation codebook provides non-contiguous positions for each pulse, such that computational complexity is reduced.

11. Apparatus as claimed in claim 6, wherein the source excitation codebook considers non-contiguous positions for each pulse, such that computational complexity is reduced.

12. Apparatus as claimed in claim 6 further comprising an optimization circuit coupled to the pitch predictor and the source excitation codebook, the optimization circuit sequentially optimizing the pitch predictor and the source excitation codebook.